We’ve commented[1] positively on IBM‘s storage and data management portfolio before. Of course, portfolio announcements, plans and strategies are all well and good. But, success depends on execution and results. Has IBM delivered? Let’s examine the results.
IBM is acknowledged as #1 in several storage-related areas, including Archiving, Storage Software apps (management, operations, collection, etc.), Mainframe Storage and branded tape. It is also #1 in Analytics. IBM states that 87% of the Fortune Global 100 companies use IBM Spectrum Storage. Overall, IBM has laid claim to being the world’s #2 storage company.
Not resting on past laurels, IBM storage continues to expand and innovate its data management and storage portfolio to resolve pressing challenges facing enterprise operations. Today’s enterprise world of multi-cloud environments coupled with on-going explosion in the amount and different types of data, poses special problems which significantly impact IT operating infrastructure, apps and services. These include escalating pressures for protection and security, speedy implementation and application of new technologies, enhancing and extending existing products and leveraging all this change for competitive advantage. Challenges that must be addressed by both vendors and their enterprise customers.
IBM plans to maintain its success in a rapidly evolving market with aggressive, fast-paced innovation aimed at new product creation, as well as enhancement and extension of the existing portfolio. To that end, IBM recently announced new, along with current and planned enhancements to existing products and services. First, we examine the current enterprise operating environment. Here’s what’s going on.
What’s the issue?
Today’s enterprise must compete in a globally dispersed, data-driven, rapidly evolving and fiercely competitive environment. This means success for today’s enterprises depends upon data, lots and lots of data. From frontline sales to DevOps, solution and service delivery, IT infrastructure and business operations, it’s the collecting, supporting, processing and reporting of data that’s on everyone’s mind. More precisely it’s the insight and innovation based on data analysis and manipulation that, in the end, delivers lasting, significant competitive advantage.
In addition, many of today’s successful enterprises have reached a consensus that the preferred infrastructure and operating environment to achieve the desired pace of agile development and delivery is a hybrid mix of private cloud (on or off-premise) plus multiple public (also on/ or off-premise) clouds. See Figure 1.
|
Enterprises today require an operations environment and infrastructure that is fast, secure, extensible, able to work transparently across multiple cloud configurations, adaptable enough to leverage next generation applications and emerging technologies, all available in an extremely cost-effective manner. One based on and adaptable to the latest technology.
In addition, operations and applications supporting data collection, management and processing services must mesh and function in a multi-cloud mix.
IBM has plans, as well as new and enhanced products and services to address these needs. Here’s what we think.
Creation to Archive – Innovation drives action
IBM’s announcement included new products (Spectrum Discover, the NVMe-enabled Storwize V7000 Gen3 & Tape drives), as well as significant enhancements to storage systems infrastructure (FlashSystem, Cloud Object, Storwize), storage services, software products for management, scaling, virtualization and cluster virtualization. Figure 2 provides a view of the breadth and depth of the products affected. Clearly, there is far too much to cover in a single document. We’ll summarize details on selected pieces that we think stand out.
 |
(Image courtesy of IBM, Inc.)
Figure 2 Map of new and enhanced storage products
|
Specifically, we will discuss end-to-end portfolio support of NVMe[2] and NVME-oF[3], the new Storwize V7000 Gen3, FlashCore enhancements, the new IBM Spectrum Discover product and Spectrum enhancements.
Infrastructure – faster, more capacity, end-to-end NVMe
IBM is no stranger to flash storage and networking solutions. They’ve consistently pushed the boundaries of speed, capacity, data protection and security. They do so again to deliver end-to-end NVMe support across the IBM storage portfolio. They are also increasing transfer rates, reducing latency and increasing data volume capacity. We provide some specifics below. But first, why is NVMe support critical?
As mentioned, the task of data storage, management and processing is an enormous part of enterprise IT operations and critical to business success. We mentioned the embrace of multi-cloud environments. As a result, rapid movement and manipulation of massive amounts of data across this mixed cloud environment must be extremely quick, easy, transparent, secure and consistent.
This is the role of NVMe and NVMe-oF connect systems and servers with the full range of networked peripheral, network fabric and bus-connected devices. To provide an idea of the scale of improvement, devices using traditional data transmission protocols can only handle 32 commands/message queue (SATA) or somewhat faster 256 commands/message queue (SAS). NVMe can handle 64,000 commands/queue and has maximum of 65,535 input/output message queues.
Without such capacities, applications and services leveraging AI, Machine Learning (ML) and the analysis of massive amounts of data would be severely limited, if possible at all.
IBM Storwize V7000 Gen3 is built to address the challenge. Storwize V7000 Gen3 system uses enhanced compression technologies and advanced AI-based management to make optimal use of NVMe internal to the Storwize array and NVMe-oF technologies. NVMe dramatically speeds data movement between systems and storage devices with much reduced latency (which remains constant at all capacity levels). It increases both the speed of transfer (16Gb NVMe over fibre channel interfaces) and the volume of data moved. Both important in today’s data-heavy environments. IBM reports throughput increases of 2.7x with 25% more available capacity in the control enclosure over current products. That translates to a maximum capacity per control enclosure of 461T and the maximum capacity in a 4-way clustered system of 32PB. This exceeds anything currently available elsewhere.
IBM FlashCore technology has also been optimized for NVMe, resulting in improved performance which translates to business advantage at reduced cost. For instance, AI-driven optimization of storage management and data placement maximizes use of time, money and resources. Knowledge of data usage patterns allows Easy Tier AI to automatically move data to the most appropriate media tier. Hardware-implemented encryption and compression means there is no negative impact on performance. FIPS 140-2 certified security with end-to-end NVMe on FlashCore Modules + SAN infrastructure meets enterprise security requirements.
In a significant move from the past operations, starting October 23, 2018, all Storwize products will be ordered through and delivered by IBM’s global network of Business Partners. This applies to all the new NVME-based Storwize V7000 Gen3, as well as earlier Storwize V7000F, V7000, V5030F, V5020 and V5010. It also applies to VeraStack systems.
IBM is not decreasing its marketing, sales and tech support of its storage line. In fact, IBM is adding Storwize sales resources to focus on the Storwize family with its channel partner community. IBM marketing, sales and technical support for the Storwize family remains unchanged. Business Partners and customers will have local support and services with the opportunity to realize greater advantages and increase benefits.
Infrastructure – Flash
The big news here is near pervasive support for NVMe and NVMe-oF. This significantly speeds the transfer of much larger amounts of data. This improves and speeds data access and accelerates application performance. NVMe-oF support is implemented via a simple, non-disruptive, no-charge IBM Spectrum Virtualize software update. This applies to FlashSystem 9100 and other members of the IBM Spectrum Virtualize family. (Some restrictions do apply, so check eligibility.)
Plans for 2019 include:
|
Performance and efficiency enhancements to the IBM FlashSystem 900 appear in reduced write latency (85µs) and a high speed (16Gb) Fibre Channel NVMe SAN interface. A new double capacity MicroLatency Module (44TB with in-line compression) doubles the effective capacity to 440TB. Increasing capacity without changing system foot print means there is no increase in power load or cooling requirements, leading to cost savings.
The IBM DS88880 all-flash Data System maintains the same footprint but uses a new 15.36 TB custom flash module delivering up to 2x boost in supported flash capacity to 8PB maximum (varies by model). Of course, Easy Tier AI-based benefits of intelligent automated data placement are also available.
A quick word on security and data protection, both increasingly critical today. Between governmental mandates, such as GDPR, HIPAA, FIPS, etc. and sophisticated global hacking, protection and security techniques must constantly evolve. IBM addresses this at multiple levels. IBM Cloud Storage provides cost-effective solutions built around policy-based WORM and lockable vaults. Cybercrime, ransomware and other mandates make secure archive and backup more critical than ever. Physically isolated tape media provide “air gap” protection for backup/archiving. IBM’s new high capacity (20TB) drive, the TS1160, has improved performance and is compatible with existing IBM TS3500 and TS4500 libraries.
 |
(Image courtesy of IBM, Inc.)
Figure 3 What’s new and enhanced in the Spectrum portfolio |
Software and Solutions
IBM announced IBM Spectrum Discover as well as enhancements to multiple IBM Spectrum products. See Figure 3. IBM Spectrum Discover allows fully automated, fast review and characterization of billions of files of unstructured data. This allows potentially useful data to be identified for further processing.
The amount of unstructured data available today is staggeringly large. To accurately classify, characterize, sort and index unstructured data for analysis has been an unbelievably time-consuming, difficult and error-prone activity even with high-powered computers and dedicated data scientists.
IBM Research provided the technology and techniques that allow IBM Spectrum Discover to automatically process metadata to classify and index billions of unstructured files and objects. It identifies data that can be used by AI, ML and analytics to yield insight and information. IBM Spectrum Discover works with unstructured data in IBM Cloud Object Storage and IBM Spectrum Scale, with support for Dell-EMC Isilon planned for 2019.
IBM also announced a number of storage and data management solution, existing and planned, which include:
- IBM Storage Solution for SAP – for data protection combining IBM storage, IBM Spectrum Protect and IBM Spectrum Copy Data Management – includes SAP TDI certification of FlashSystem 9100, the new IBM Storwize V7000 and Elastic Storage Server
- IBM FlashSystem 9100 – received certification from Epic for electronic healthcare records (note Meditech certification is coming in 2019)
- Support for blockchain – planned for 2019
- IBM Spectrum Virtualize for Public Cloud – Amazon AWS availability is planned for 2019
We’ve only covered a portion of the announcement. This should give you a good flavor of the depth and breadth of IBM’s commitment and aggressive advance in this market.
Conclusion
IBM’s describes its strategy as focused on “Innovating from Creation to Archive” as they expand and enhance an already extensive family of product, service and infrastructure offerings. On first hearing, that appears to be a bit of bragging. After the briefing and reviewing the deliverables and planned portfolio additions and direction, it is not a hollow brag. There is solid substance to what they have delivered and are planning.
We must admit that we came away impressed with what IBM has delivered. They did not just maintain, but significantly improved their portfolio. Congratulations to them.
Finally, there is much more in IBM’s announcement, including usage-based pricing, improved scale-out capabilities, improved, cost-effective archiving as well as new high-availability and disaster recovery entry configurations for lower CAPEX. We highly recommend a conversation with IBM to any enterprise with hi-volume, large capacity, fast response data and stotage needs [4].
[1] https://ptakassociates.blogspot.com/2017/11/ibm-spectrum-with-cluster.html
[2] NVMe = Non-volatile Memory Express
[3] NVMe-oF = Non-volatile Memory Express over fibre
